SWER Broken

Conductor Fault
Detector R&D Project
Stage 1- Final Report for Department of
Environment, Land, Water and Planning (DELWP)
Report Summary
18.07.19
Document Control and Non-Disclosure Notice

Revision Description Authors Approver

1.0 For DELWP website release Cagil Ozansoy (VU) & David Wilkinson (UE) Dave Denny (UE)

The information in this document has been provided for the Department of Environment, Land, Water and Planning
(DELWP) to summarise the work undertaken in the SWER Broken Conductor Fault Detector R&D project.

Victorian government funding (via DELWP) has been directed into the development of bushfire mitigation technology for
the benefit of the Victorian community. United Energy (UE) provides consent for this document to be published by
DELWP in the interests of transparency.

The information contained in this report is provided in good faith. No representation, warranty or undertaking,
expressed or implied is or will be made or given and no responsibility or liability is or will be accepted by United Energy
and Victoria University or by any of its directors, employees or advisors in relation to the accuracy or completeness of
this document or any other written or oral information made available in connection with this project by United Energy
and Victoria University. The contents of this document have not been independently verified.

This document is not to be reproduced, disseminated, copied, disclosed, modified, distributed and or published by any
other entity other than DELWP without the written consent United Energy. All Intellectual Property (IP) gained from this
project remains the property of United Energy.

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 2

Project Team

Victoria University and United Energy would like to acknowledge the important contribution to the success of this
project by the following people and organisations:

Name Organisation Name Organisation

Dr Cagil Ozansoy Victoria University Mr. David Wilkinson United Energy
Dr Mohd Noor Islam Victoria University Mr. Dave Denny United Energy
Professor Mike Faulkner Victoria University Mr. Thusara Perera United Energy

Mr Douglas Gomes Victoria University Mr. Robert Doyle United Energy

Mr. Mahamudul Hasan Victoria University Mr. Stephen Watt Zinfra

Mr. Miroslav Radev Victoria University

The project team acknowledges in-kind contributions by Energy Safe Victoria (ESV).

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 3

Executive Summary

Single Wire Earth Return (SWER) distribution systems are commonly used in Australia, New Zealand and some other
parts of the world for the electrification of rural areas. A single wire (or electrical conductor) is used to supply electrical
power with the return current flowing through the ground. SWER networks consist of thin and typically long conductors
under high tension. Under certain fault conditions the SWER conductor is prone to breaking, e.g. when a tree falls over
the powerline on a windy day. This is likely to result in the broken SWER conductor falling to the ground or into
vegetation and creating a phase to earth short circuit fault. These faults have a high risk of starting a bushfire.

SWER networks are typically installed in rural areas with more than 28,000km installed in Extreme and High Bushfire
Risk Areas (HBRA) across Victoria. The “SWER Broken Conductor Fault Detector R&D Project” aims to develop a
communications-assisted scheme to detect broken conductors on SWER powerlines. The proposed scheme relies on
continuous transmission of test signals from one end of the line and listening to these test signals at a receiver placed at
each SWER automatic circuit recloser (ACR). If the signal is lost, then the ACR will be instantaneously tripped,
interrupting the electricity supply prior to it having an opportunity to start a bushfire.

The following key tasks were undertaken as part of stage one of this project:

Literature Review: This task explored powerline carrier (PLC) application in power networks. Modeling of a power
system network as a communication channel was studied. Communication interference (electrical noise), signal
attenuation and distortion, and different PLC standards were examined. Different digital transmission and modulation
schemes were reviewed.

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 4

 Executive Summary

SWER Background Noise Analysis: This task focused on development of a hardware set-up for recording background
electrical noise levels over a SWER communication channel. The developed hardware was connected to the network
and noise recordings were taken for both High Frequency (HF) and Low Frequency (LF) channels in continuous and
sweep modes. Data was analysed to identify levels and potential sources of noise.

Network Coupling Design: This task focused on network data collection and design of a number of coupling methods.
Different matching circuits were built and signal strengths/attenuation measured.

Injection and Modulation Analysis: This task compared different injection coupling methods. Injection tests were
undertaken on three separate days. These include offline (isolated network) and online (HV connected) injection tests.

Prototype Development: This task focused on the development of a prototype fault detector inclusive of a transmitter
and a receiver. An experimental LV injection test setup was successfully constructed in the Victoria University (VU)
laboratory and used to evaluate the developed system under different configurations.

Key Conclusion:

A PLC based SWER broken conductor protection system is feasible. The optimum operating frequency range was
identified with sections of the spectrum providing adequate noise floor to achieve a high signal-to-noise ratio. Further
development is recommended.

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 5

Project Background: What is SWER?

A SWER network is a low cost solution for the supply of small amounts of electrical power in sparsely populated rural
areas. It is used widely throughout regional Australia in all states and is used in New Zealand. The design is low cost
because it uses just a single electrical conductor (typically zinc or aluminium coated steel) unlike other designs that
use two or three wires. The single conductor may stretch for tens or even hundreds of kilometres (supported by
poles), with distribution transformers tapped off to provide the low voltage (LV) supply to customers along its length.
The distinguishing feature is that the ground (or earth) becomes the return path for the current.

For two and three wire systems any substantial current flow to earth is abnormal and can be detected as a fault
allowing protection systems to operate to disconnect the electricity supply. But for the SWER network the normal load
current is designed to flow to earth so this method cannot be used to detect faults.

SWER protection is therefore generally less sensitive at detecting faults than the protection on two and three wire
systems.

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 6

Project Background: What is SWER?

The SWER design simplicity offers some advantages. With a single wire there is no possibility that multiple electrical
conductors can clash together under the influence of wind to create a fault. There are fewer electrical assets and less
that can go wrong. Fault currents are lower and less energy is released when a short circuit fault occurs. As a result
SWER powerlines are generally more reliable than two and three wire systems and more reliable systems start fewer
electrical fires.

However SWER networks consist of thin and typically long conductors under high tension and the conductor is prone to
breaking if hit by a falling structure, such as a tree. Although good design standards, workmanship and general asset
management including asset inspection and replacement programs can reduce the risk, the steel conductor is still prone
to failure.

Some SWER conductors fall to the ground and start fires. Some have resulted in catastrophic bushfires such as those in
Victoria on Black Saturday in 2009.

On days of greatest bushfire risk the wind speeds are high increasing the risk of electrical conductor failure.
Furthermore vegetation and soil is typically very dry. A broken SWER conductor resting on the ground or vegetation may
only cause a few amperes of current to flow (with high impedance fault path under dry conditions) and existing protection
may be slow or may not operate at all. But sufficient current could still flow to result in fire. It is often a challenge to
detect such high impedance phase to earth faults. The fault current is similar to the normal load current.

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 7

Project Background: What is SWER?

• SWER = Single Wire Earth Return

Current supplied by a single wire

Current returns in the ground

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 8

Project Background: What is SWER?
.7 kV
......... .........

22 kV

12.7kV/240 V 12.7kV/240 V
Local Power Local Power
Transformer Transformer Remote-controlled
Automatic Circuit
Recloser

Typically 300kVA
22/12.7 kV
ACR Control and
Isolation
Communications
Transformer
Cubicle

. . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . .
Earth return path Earth return path

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 9

Project Background: Project Genesis and Drivers
On Black Saturday, 7th February 2009, a SWER conductor failure started a catastrophic bushfire in Kilmore East. It
was this event which is the driver behind the need to develop new technology to reduce the risk of SWER powerlines
starting bushfires.

While new SWER ACRs have been deployed with special total fire ban (TFB) operating modes, overhead lines have
been replaced by underground cables and new covered conductor (insulated overhead) lines have been built in
extreme risk areas, the majority of SWER powerlines use a bare overhead conductor and there is a risk of a SWER
conductor falling to the ground and starting a bushfire.

United Energy has proposed and initiated the development of broken conductor fault detection technology for the
benefit of the Victorian and Australian community.

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 10

Project Background: Aims and Objectives

Even if the development of the SWER broken conductor fault detector is successful, and it is widely adopted, it
will not prevent all SWER powerline faults from starting bushfires. It may not even prevent the majority. Not all
faults and fires that are started from SWER powerlines result in the break of a conductor. However the
technology is expected to provide an additional control to reduce risk that is affordable and when combined with
other technologies such as low set instantaneous overcurrent protection and disabling of reclose will reduce risk
to a level that is comparable to multi-phase REFCL protected lines. If R&D continues it is also possible the
technology could be improved via low cost firmware updates (no new hardware) to add fault signature detection
algorithms that may expand upon the type of faults that can be detected.

The overall objective of stage one of this project was to assess the feasibility of a PLC communications based
SWER protection scheme in the detection of conductor breakage in SWER networks. The specific aims in stage
one were:

• The measurement and recording of background electrical noise over the PLC frequency range of interest.
This background noise has potential to interfere with the ability of the receiver to detect the PLC signals
transmitted.
• Analysis of background electrical noise to determine the optimum frequency bands over which to operate,
the required transmitter power levels and the modulation and signal encoding methods to use.

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 11

Project Background: Aims and Objectives

• Detailed desktop review of PLC communication technologies and their use in protection applications.
• Tests to evaluate different techniques in regards to different transmission frequencies, modulation methods,
error correction techniques, communication time delays, signal to noise ratios, impact of intermittent
interference and high power signal jammers.
• Design and analysis of three different transmitter network coupling methods.
• Investigation into electromagnetic compliance issues both in regards to compliance with regulations and the
impact the transmitted signals have on the operation of other equipment.
• Evaluation of the need for line traps.
• Development of a prototype fault detector ready for deployment on a real SWER network with multiple
transmitters and a single receiver.
• Material to support the continuation of R&D for stage two including cost-benefit analysis as required to
prepare a business case.
• To partner with a Victorian based University to support local innovation and foster academic / industry
collaboration. (Victoria University was selected following a competitive selection process.)

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 12

Project Background: PLC-Based SWER Protection

The concept relies on detecting broken conductors by placing PLC transmitters at certain positions at the end of
SWER lines with a receiver at each ACR. If the signal is lost, then the ACR is instantaneously tripped.
Coupling Coupling
Capacitor Capacitor

12.7 kV

2
Coupling
Device
1
Line Trap

Coupling 22 kV
Device

Remote-controlled
Automatic Circuit
12.7kV/240 V Recloser
Local Power
Isolation
Transformer
Transformer

ACR Control and

Communications
Cubicle

3
PLC PLC
Transmitter Receiver

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 13

Project Background: Myers Road SWER Network – test site

The “Myers Road” SWER network is located

in Balnarring, Victoria. The network is a
relatively small SWER network with a total
network length of 10.2 km.

The Myers Road SWER network serves 28

customers all supplied power through the
12.7 kV SWER network.

SWER Network
Data logging and tests were undertaken at
the “Myers Road” isolation transformer
(ISO). Note the ISO isolates the SWER
network from the 22kV multi-phase network,
changing the voltage from 22kV to the
SWER voltage (12.7 kV).

“Myers Road” SWER network in Balnarring.

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 14

Project Background: Myers Road SWER Network

The SWER ACR is located on a second pole approximately 22 meters away from the pole where the test set up was
located.
The data logging test equipment was located in an enclosure placed 3 meters high on Pole #1.

Isolation Transformer
SWER conductor
SWER ACR

Pole #2
Test Set Up

Pole #1

“Myers Road” SWER ISO and ACR.

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 15

Project Background: Noise Logging
Target Bandwidth = 10 kHz to 500 kHz | 0 Hz to 50 kHz
Key elements of the background noise logging Suitable
Conductor
Receiver Noise Analysis Concept Design
included the: 1.1 nF coupling capacitor, a 12.7 kV HF Channel
Low-Noise Pre-Amplifier/Filter
coupler termination box, a custom designed and MCC124 Coupling
Digitiser Transmitter
Sampling @ 5 MS/s
Capacitor 1.1nF
Low-Pass Bessel IIR AA Filter < 500 kHz
built Capacitive Voltage Divider (CVD), a low- 12.7 kV
RG58
Coaxial
noise filter, two Digitisers and the Gen 3i logger. Second-stage
RG58
Coaxial 1m
over-voltage
0.5m
protection

The objective was to develop a logging concept TNC HF Digitiser

CVD HP F/O
that enabled voltage background-noise data BNC BNC
BNC BNC
collection and analysis to be carried out for the 11 m
RG58 RG58 Input LP
Coaxial Gen 3i
communication channel. Coaxial
0.5m Logger

RG58
OSC 1m
Coaxial
CTB1A
The two CVD channels were recorded Power GND
5m RG58 F/O
Diagnostix Coaxial
individually. This enabled background noise Systems Coupler
Termination Box
BNC
levels of the SWER network to be measured in
LF Digitiser
two channels: the 10 kHz to 500 kHz High
BNC LF Channel
Frequency (HF) channel and the 0 Hz to 50 kHz Digitiser Transmitter
Sampling @ 5 MS/s
Low-Frequency (LF) Channel Low-Pass Bessel IIR AA Filter < 50 kHz

“Myers Road” SWER network noise analysis concept.

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 16

Project Background: Transmitter Setup

The transmitter set-up was connected

at the “French No: 3 Bittern” SWER Coupling Capacitor
distribution substation.
The pole is located next to a
farmhouse and an enclosure was
erected 3 meters high on the pole SWER Conductor
SWER
enabling the placement of all relevant Transformer
injection hardware.
Transmitter

The direct distance between the Enclosure

transmitter and receiver is

approximately 3.6 km.

A coupling capacitor was mounted on

the pole enabling the injection of HF
signals directly onto the 12.7 kV
conductor through the coupling
capacitor. “French No: 3 Bittern” SWER Distribution Transformer Pole.

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 17

Project Background: Transmitter Setup

Three different injection options SWER TR Low Voltage Winding Injection and Coupling Circuit
were designed and tested. These 12.7 kV SWER conductor
Electrical Plug RG58
include the coupling capacitor, MV RG58 4 Amps Fuse
Coupling
Circuit
earth, and SWER Transformer Low RG58
CTB1A
GPO
Voltage Winding injection A
RG58
approaches. RG58 Signal
Generator
Filter

Coupling Circuit

The SWER substation LVW circuit RG58

Toroidal N 80 Amps Fuse
CT
included the matching circuit, a Filter
PLC Filter
Customer
low-pass filter, and the arbitrary Load
RG58
signal generator, and it was
connected directly to a 240V Signal MV Earth LV Earth
Generator
general purpose outlet (GPO) RG58 MV Earth Injection and Coupling Circuit
located within the enclosure. The Coupling Capacitor
Injection and Coupling Coupling RG58 RG58
Signal
MV Earth injection circuit included Circuit Circuit
Filter
Generator
a Toroidal Transformer.

Injection Set-Up at “French No: 3 Bittern” SWER Distribution Substation

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 18

 Project Deliverables – Literature review

There are limited published applications of real world distribution PLC

protection applications. Published academic papers focus on abstract
concepts.

Selection of the optimum frequency range for PLC signal transmission is

critical. Background noise/signal analysis over the communication channel
should be performed before the selection of the appropriate frequency.
There is a need to balance between signal attenuation, signal to noise
ratio (SNR) and compliance with regulations.

The type of modulation and error correction will have significant impact on
the signal-to-noise ratio and accurate transfer of any key information.

No mandatory standard for the Electromagnetic Compatibility (EMC) for

HV distribution systems exists. The CENELEC Standard EN 50065 (for LV
PLC installations) has been used as a guide.

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 19

Project Deliverables – Noise Measurements

Communication over powerlines is

degraded by external electrical noise, and Pole #1
the signal to noise ratio (SNR) plays a key
role in the reliability of a PLC scheme. Pole #2
SWER conductor

Specialised electrical circuits were custom

SWER ACR
designed and built to record noise with the Isolation Transformer
network in energised and de-energized
states.

The test equipment was located in an

enclosure placed 3 meters high on Pole #1 Test Set Up
at the “Myers Road” SWER ISO.

The noise recordings were analysed to

understand the impact on the ability to
successfully communicate using PLC
signals.

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 20

Project Deliverables – Noise Measurements

Based on the investigation of the SWER

communication channel background noise Pole #1
levels, a PLC application is feasible.
Pole #2
In the frequency range of interest, sections SWER conductor

of the spectrum have sufficiently low noise

SWER ACR
to permit reliable communication. Isolation Transformer

The noise floor levels are sufficiently low

enough to achieve a high signal-to-noise
ratio even for received signals as low as -
60 dB. Test Set Up

This would ensure a minimum signal-to-

noise ratio of 10 dB, which is sufficient for
the reliable operation and decision making
in a receiver.

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 21

Project Deliverables – Coupling Circuits

Three injection coupling methods were modelled including the Coupling Capacitor (CC), Low Voltage Winding (LVW),
and Medium Voltage Earth (MVE).

Coupling Capacitor (CC) Medium Voltage Earth (MVE) Low-Voltage Winding (LVW)
Injection Circuit Injection Circuit Injection Circuit

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 22

Project Deliverables – Injection Tests

Signal strength received for the three signal injection methods were compared over a range of frequencies. The LV line
injection method performed the best and is a low cost option. The SWER distribution transformer LV winding coupling
has been selected as the signal injection method for the prototype fault detector.

Coupling Capacitor SWER TR Low Voltage Winding Injection and Coupling Circuit
12.7 kV SWER conductor
Electrical Plug Coupling
RG58
RG58 4 Amps Fuse Circuit
SWER Transformer RG58
CTB1A
GPO
A
RG58
SWER Conductor RG58 Signal
Filter
Generator
Coupling Circuit A

RG58 80 Amps Fuse

Toroidal N
CT
PLC Filter
Filter Customer
Transmitter Enclosure Load
RG58

Signal MV Earth LV Earth

Generator

RG58 MV Earth Injection and Coupling Circuit

Coupling Capacitor
Injection and Coupling Coupling RG58 RG58
Signal
Circuit Filter
Circuit Generator

Pole #1
Injection Set-Up at “French No: 3 Bittern” SWER
Distribution Transformer.
“French No: 3 Bittern SWER Distribution Transformer Pole.

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 23

Project Deliverables – Injection Tests
A summary of test days and tested injection methods are provided below.

Test Date Method Mode Comments

7th of August 2018 Coupling Capacitor Online & Offline Successful injection during offline periods
17th of August 2018 Coupling Capacitor Online Successful injection
14th of December 2018 Coupling Capacitor, Medium Voltage Earth, SWER Online Successful injection
Transformer LV Winding

On the 7th of August 2018, successful injection was achieved when the Isolation Transformer was isolated from the 22
kV three-phase network. The failure of a Low-Pass (LP) filter in the injection set up prevented successful injection
during the online injection periods. This necessitated the online tests to be repeated on the 17th of August 2018, when
the coupling capacitor based injections were successfully retested for the connected Myers Road SWER network.
On the 14th of December, all three-injection methods were tested under the same weather and network conditions in
the 22 kV grid connected network. These include the Coupling Capacitor (CC), Medium Voltage Earth (MVE), SWER
Transformer Low Voltage Winding (LVW) injection methods.
SWER Transformer LVW based injection was observed to be superior to the other two methods in terms of both
received signal strengths and peak correlation levels. This is a key advantage in the direction towards cost-effective
implementation of the proposed scheme without requiring the use of expensive coupling capacitors or CTs.

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 24

Project Deliverables – Prototype Development

The PLC prototype consists of a transmitter and a receiver. A combination of Digital Signal Processor (DSP), ARM
microcontroller and Field Programmable Gate Array (FPGA) have been used in developing the prototype.

Transmitter Prototype. Receiver Prototype.

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 25

Project Deliverables – Prototype Development

An experimental LV injection test setup has been successfully constructed in the VU laboratory.
The experimental setup was used to test the developed system under different configurations including
(i) noiseless modulated signal transmission, (ii) recordings of the “Myers Road” SWER network background noise and
(iii) recording of a transmission test on the network including the PLC modulated signal and background noise.
Results show that the autocorrelation function implemented in the Receiver can successfully identify the transmitted
signal even in cases of strong real background noise.

Laboratory experimental setup.

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 26

Stage 1 Project Findings

Various digital modulation, error correction techniques, and suitable communication protocols were reviewed as part of
the literature review. Different PLC transmitter & receiver coupling techniques were analysed.

Background network voltage noise recordings were successfully gathered over a six-day period, dominated by phase-to-
ground voltages. This includes ten-second continuous mode recordings taken every 15-minutes for a period of five days
and trigger based sweep mode recordings for three days.

Peak-to-peak periodic pulses were observed at the HF channel associated with Variable Frequency Drive (VFD) equipped
electro-mechanical systems.

A noise floor of -70 dB was observed in the designated bandwidth, ideal for signal transmission requiring -60 dB received
signal level for a practical SNR of +10 dB.

The direct path between a transmitter (TX) and receiver (RX) was observed to be most critical in signal attenuation with
the lengths of adjoining spurs having minimal impact.

In injection tests, the LVW coupling based injection produced signal strengths and peak correlation levels superior to the
other two methods. This is expected to aid in the development of a cost-effective solution as SWER transformers are
readily available and connected on the network.

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 27

Conclusion

A collaborative R&D project between UE, DELWP and VU has enabled a prototype SWER broken conductor fault detector
to be developed that can rapidly detect the break in a SWER powerline and de-energise the powerline before the fault has
an opportunity to start a bushfire.

Simulations and testing undertaken in the laboratory have demonstrated that the prototype is expected to perform well on
a real SWER network, however the technology remains immature and the prototype is yet to be installed and tested on a
real SWER network.

Further development is proposed to install the prototype protection system on real SWER networks to monitor and
improve upon the design. Challenges remain including the impact of communication interference, PLC signal losses
associated with transitions from overhead and underground cable, spur fusing and others however solutions may also be
available.

If testing on real networks is successful and the technical challenges can be overcome, the final stage will be to develop a
mature robust, reliable and cost effective commercial product.

The ultimate objective is to develop technology that can be widely deployed to reduce the number of bushfires started
from SWER powerline in Victoria and beyond.

18.07.19 SWER Broken Conductor Detector R&D Project - Final Report 28